PREDICTION OF SUNFLOWER CROP YIELD USING COMPUTER SOFTWARE APPLICATION

Agricultural yield prediction procedures are often proposed to explore for techniques or models that discover the practical relationship between influencing variables and production. In this research an Aadaptive Neuro-Fuzzy Inference System (ANFIS) was used to build a model to predict sunflower crop yield grown in Egypt. The inputs to ANFIS were crop head diameter, seed numbers in the head, seed moisture content and pre-harvest losses. The output was sunflower seed crop yield. ANFIS membership function the triangle generally gave the most desired results with respect to mean error, root mean square error and coefficient of determination (R2) statistical performance testing tools. The results revealed that ANFIS model was able to predict sunflower yield with a satisfactory performance with mean error of 18 kg/fed in testing phase. Coefficient of determination between ANFIS output and actual yield was 0.996 for testing data. However, the mean error and R2 values were 33.8 kg/fed and 0.702 when using multiple linear regression model in yield prediction. The results demonstrate that ANFIS can be applied successfully and provide high accuracy and reliability for sunflower crop yield prediction. The findings of this research could be used as a crop management tool.